Symposium 6: Memory and Amnesia in Man and Animals

behaviour as a result of experience. It is a prerequisite for episodic memory, which is selectively impaired in clinical amnesia. Primate models of amnesia are valid only if monkeys possess the capacity for secondary mental representation and only if the tests employed depend critically on its use. It is not self-evident that the conventional test of amnesia in monkeys (impaired performance on delayed non-matching to sample tasks) does depend on secondary mental representation. Secondary mental representation is also necessary for the comparison of alternatives within one cognitive manoeuvre. Monkeys with hippocampal dysfunction are impaired on conditional learning when trials are presented in an order which invites comparison of alternatives, but are not impaired on the same task when the order of trials artificially segregates component parts of the task. The hippocampus may be particularly important in the process of secondary mental representation, and it is this contribution to cognition which may underlie the critical involvement of the hippocampus in episodic memory. Rather than seek to model amnesia directly in rats, we have adopted the approach of analysing component aspects of memory and uncovering the neural systems supporting these various elements. Particular attention has focussed on functional relationships with the hippocampus as this structure is critical for normal episodic memory in humans. Two techniques have been used to uncover functional relationships among regions that might play a key role in memory. Lesion studies have highlighted the many similarities between the effects of damage to the fomix, anterior thalamic nuclei, and hippocampus on tests of spatial memory; similarities that have been supported by recent disconnection studies involving contralateral lesions of the anterior thalamus and fomix. In contrast, regions such as the perirhinal cortex and the medial dorsal thalamic nucleus show quite different lesion induced deficits that can be dissociated from those of the hippocampus. Mapping studies using the expression of the immediate early gene c-fos support these distinctions. Perihinal cortex shows increased gene activity when a rat is shown novel visual items, while the hippocampus remains unchanged. In contrast, the hippocampus and anterior thalamic nuclei, but not the perirhinal cortex, show enhanced fos production after performing allocentric spatial memory tasks. Importantly, there are changes in c-fos activation in both the hippocampus and perirhinal cortex when the spatial configuration of familiar items is changed. Such findings not only highlight the importance of the anterior thalamic nuclei for hippocampal function, but also reveal a …