Our overall hypothesis is that the amygdala is specifically
important for storing associations with the intrinsic incentive
value of a reward event, and not with other attributes of the
event. In experiments with secondary reinforcers, this type of
association is formed between the sensory properties of the secondary
reinforcer and the incentive value of the primary reinforcer. Once formed, that association sustains the secondary
reinforcer’s power to evoke emotional and other reactions, the
third type of attribute of a reward event enumerated above.
When, as in the 2 tasks considered above, a visual discriminative
stimulus is followed by a secondary reinforcer, there are 2 possible
routes for discriminative learning: The discriminative
stimulus may enter into association either with the emotional
reactions aroused by the secondary reinforcer or with the reinforcer’s
sensory properties. The amygdala is important for mediating
the learned emotional reactions to the secondary reinforcer.
Therefore, when discriminative learning is dependent on
the first of these 2 processes (as in the task of Gaffan and Harrison,
1987), amygdalectomy produces a large impairment, but
when the second, sensory-sensory type of association predominates
(as in the present task), this surgery has less effect.