Although certain neurophysiological functions of the amygdala complex in learning seem well established, the purpose of this review is to propose that an additional conceptualization of amygdala function is now needed. The research we review provides evidence that a subsystem within the amygdala provides a coordinated regulation of attentional processes. An important aspect of this additional neuropsychology of the amygdala is that it may aid in understanding the importance of connections between the amygdala and other neural systems in information processing.Kluver and Bucy (1) Perhaps the most frequently cited measure of a cue's changed motivational value, as a consequence of its pairing with a biologically meaningful stimulus, such as food, is its ability to reinforce subsequent learning-for example, the phenomenon known as second-order conditioning. We tested a cue's ability to support Pavlovian second-order conditioning in rats with CN damage. As shown in Fig. 1, second-order conditioning was as robust in rats with CN lesions as it was in normal control rats (see legend to Fig. 1 for description of task).Thus, CN damage did not interfere with a cue's ability to acquire reinforcing power when paired with food.We also tested whether a cue's acquisition of motivational properties remains intact in rats with CN damage by using a procedure called "conditioned potentiation of feeding." In this procedure, sated rats are induced to eat by presenting a previously trained cue for food (9). To test for this effect, rats were placed back on ad libitum feeding after standard appetitive training and exposed repeatedly to the test chamber with food available to eliminate all tendency to consume food in that environment. Two test sessions were then conducted, during which 20 food pellets were made available. In one session the formerly trained cue was presented; in the other no cue presentations occurred during the test. Our experiment showed that feeding was potentiated substantially by the cue in normal rats; whereas an average of 0.2 pellet was consumed in the unsignaled test, on average, 15.2 pellets were consumed in the cue-signaled condition. The rats with CN damage showed a comparable effect of cue presentations; lesioned rats ate 3.8 pellets in the unsignaled condition and 16.9 pellets in the signaled condition (10).These results provide strong support for the conclusion that the motivational value associated with cues during appetitive learning is acquired normally by rats with amygdala CN damage. Taken together with research conducted in other laboratories (e.g., refs. 6 and 8), it appears that a separate subsystem within the amygdala is necessary for this function. Despite the normal appearance of rats with CN damage on the tests described above, the lesions were highly effective in producing other behavioral impairments. One that can be observed during simple appetitive conditioning will be described here; others will be introduced in a later section of the paper.We observed that neurotoxic damage to t...