Normal olfactory discrimination learning set and facilitation of reversal learning after medial-temporal damage in rats: Implications for an account of preserved learning abilities in amnesia

Recent evidence of preserved skill learning in patients with 'global' amnesia has led to the postulation of a qualitative distinction between functionally separate memory systems, one of which may remain preserved when the other is profoundly impaired. On one account, the separate memory systems support either the learning of declarative knowledge, i.e., facts and associations, or the learning of procedural knowledge i.e., knowledge that permits the expression of skilled performance without reference to specific facts or associations. In an effort to develop a rodent model of amnesia that illustrates the same distinction between memory systems, rats were trained in a series of discrimination and reversal problems using olfaction, a sensory modality in which they rapidly learn new associations. Rats with bilateral fornix, amygdala, or combined fornix and amygdala damage learned successive two-odor discriminations as quickly as normal and sham-operated control subjects. Furthermore, all groups rapidly acquired the skills of discrimination as revealed in the development of a learning set. Subsequent presentation of a reversal of one discrimination elicited a marked dissociation among groups: Normal rats and rats with amygdala lesions required many more trials to acquire the reversal than to acquire a new discrimination problem, whereas rats with fornix lesions learned the reversal rather easily. A detailed analysis of response strategies suggested that normal rats and rats with amygdala lesions first extinguished the prior response tendencies and then abandoned the learning set skills and treated the reversal much as they did the initial discrimination problem. In contrast, after prior response tendencies were extinguished, rats with fornix lesions were relatively unaffected by the prior cue-associations and treated the reversal much as they would a new discrimination problem. These results distinguish between impaired and spared memory capacities in rodents with amnesia, in conformance with findings on human amnesia.