Howard Eichenbaum

The hippocampal system and declarative memory in animals

Theoretical arguments and empirical data are presented in favor of the hypothesis that the hippocampal system supports a declarative memory capacity in animals as well as humans. This view is advanced by identifying two prominent characteristics of human declarative memory and by operationalizing and evaluating them using both experimental lesion and single unit recording studies on animals. First, hippocampal processing is not selective to any particular category of learning materials; instead, it supports comparisons among all kinds of information in memory, resulting in a representation of critical relations between items. Conversely, individual representations are supported outside the hippocampal system. Second, hippocampal-dependent, relational memory representations involve a flexible organization that permits inferences from memory in novel situations. Conversely, hippocampal-independent individual representations can support only repetition of procedures acquired during original learning. Correspondences between the neuropsychological and neurophysiological findings presented serve to indicate how these properties of hippocampal representation support declarative memory across behavioral paradigms and across species.

Critical role of the parahippocampal region for paired-associate learning in rats.

Human amnesics are severely impaired in learning arbitrary associations between word pairs. This report examines the role of the hippocampal system in an analog of paired-associate learning developed for rats. Subjects with selective damage to the parahippocampal region and controls were tested on an odor-guided paired-associate task. Normal rats readily learned to distinguish paired associates from various mispairings of the same odors, whereas most animals with parahippocampal damage failed to learn. Lesioned animals were, however, able to acquire correct responses to stimulus pairs for which an association between odors was not required. These findings extend the usefulness of olfactory learning as a model for studies of higher order learning capacity in rodents and indicate that in rats, as in humans, hippocampal areas are critical for learning arbitrary nonspatial relationships between stimuli.

Odor-guided learning and memory in rats: is it ‘special’?

the first 20 trials of each problem after having learned no more than four such problems previously. Rapid progressive improvement is not a special feature of olfactory learning. There remains one feature of olfactory learning that might prompt us to allow it some special status, and ironically this is found in the work of Reid and Morris 3 themselves. It is their finding that serial reversal improvement fails to occur. Such improve- ment emerges quite readily when rats are trained on an analogous visual discrimination 3. One reason for this improvement, it has been suggested, is that the rats come to restrict attention to the relevant stimuli, to abandon position habits, and so on, an interpret- ation that makes Reid and Morriss failure to find the effect all the more puzzling since it is just these processes that they suggest might generate the improvement they observe in the olfactory learning-set procedure. I have no solution to this enigma. All I can do is express the hope that further research will clarify the issue of what conditions are necessary for serial reversal improvement in ol- factory learning; also, given the current state of confusion on this issue, to urge caution about taking the Reid and Morris result as a convincing demon- stration that olfactory learning is in some way special.

Olfactory Learning and Memory in the Rat: A “Model System” for Studies of the Neurobiology of Memory

The study of olfactory learning and memory in the rat not only is an interesting research area in its own right, but additionally, and perhaps more importantly, provides a means by which a comprehensive, biologically based theory of memory can be developed. The purpose of this chapter is to review the behavioral, anatomical, and physiological data supporting this view, and in doing so, to determine the extent to which the rodent olfactory system serves as a model that can be used to extend our understanding of the neurobiological bases of human memory.

The Hippocampus and the Sense of Smell

Brodal, in a 1947 review from which we boldly borrow the title of the present paper, outlined some of the critical evidence that ultimately led to the demise of the notion that the hippocampus was a part of the olfactory brain, or “rhinencephalon” as it was called according to the prevailing view of the time. Since then it has become abundantly clear that the hippocampus processes information from many input sources (cf. Deacon et al., 1983). Nevertheless, converging data from neuroanatomical, physiological, and behavioral studies indicate that the olfactory system projects heavily onto and has especially immediate access to the hippocampal system, suggesting that the olfactory-hippocampal pathway may be particularly useful for explorations of sensory-limbic interactions leading to the higher order coding of perceptual information. As will be described below, the intimate anatomical associations between the olfactory and hippocampal systems are paralleled by 1) the critical role played by the hippocampal system in odor-guided learning and memory, 2) the strong influence of olfactory processing over the physiological activity in the hippocampus both at the level of rhythmic EEG activity and at the level of neuronal firing patterns, and 3) the role these physiological processes may play in the induction of synaptic plasticity supporting memory formation. Thus, in the spirit of a “renaissance of the rhinencephalon” (Macrides, 1977), we will argue that olfaction is a particularly advantageous model system for studies of “sensory” processing by the hippocampus across behavioral, neuronal, and synaptic levels of analysis. Our data on studies at each of these levels of analysis will be discussed in turn (see also Otto and Eichenbaum, 1992b).