The authors draw together the results of a series of detailed computational studies and show how they are contributing to the development of a theory of hippocampal function. A new part of the theory introduced here is a quantitative analysis of how backprojections from the hippocampus to the neocortex could lead to the recall of recent memories. The theory is then compared with other theories of hippocampal function. First, what is computed by the hippocampus is considered. The hypothesis the authors advocate, on the basis of the effects of damage to the hippocampus and neuronal activity recorded in it, is that it is involved in the formation of new memories by acting as an intermediate-term buffer store for information about episodes, particularly for spatial, but probably also for some nonspatial, information. The authors analyze how the hippocampus could perform this function, by producing a computational theory of how it operates, based on neuroanatomical and neurophysiological information about the different neuronal systems con-tained within the hippocampus. Key hypotheses are that the CA3 pyramidal cells operate as a single autoassociation network to store new episodic information as it arrives via a number of specialized preprocessing stages from many association areas of the cerebral cortex, and that the dentate

The information represented in the primate hippocampus is being analysed by making recordings in monkeys actively walking in the laboratory. In a sample of 352 cells recorded in this situation, no "place" cells have so far been found. Instead, we have found a considerable population of "spatial view" cells tuned to respond when the monkey looks at small parts of the environment. We have been able to demonstrate (1) that these hippocampal neurons respond to a view of space "out there," not to the place where the monkey is; (2) that the responses depend on where the monkey is looking, by measuring eye position; (3) that the responses in some cases (e.g., CA1 but not CA3) still occur if the view details are obscured with curtains; (4) that the cells (in, e.g., CA1) retain part of their "space" tuning even in complete darkness, for several minutes; and (5) that the spatial representation is allocentric. The spatial representation is, thus, different from that in the rat hippocampus, in which place cells respond based on where the rat is located. The representation is also different from that described in the parietal cortex, where neurons respond in egocentric coordinates. This representation of space "out there" provided by primate spatial view cells would be an appropriate part of a memory system involved in memories of particular events or episodes, for example, of where in an environment an object was seen. Spatial view cells (in conjunction with whole body motion cells in the primate hippocampus, and head direction cells in the primate presubiculum) would also be useful as part of a spatial navigation system, for which they would provide a memory component. Hippocampus 1999;9:467--480. # 1999 Wiley-Liss, Inc.

The firing rate of hippocampal neurons in rats was related both to spatial location and to multiple behavioral variables as rats performed 2 kinds of tasks that rely on hippocampal function: a spatial navigation task similar in performance demands to the radial-arm maze task and a simultaneous-cue odor-discrimination task. In the place task, most cells had distinct single or multiple place fields, that is, neurons increased firing when the rat was in a particular location or locations. However, in most of these cells, firing rate also varied systematically in relation to behavioral variables, in-cluding the speed, direction, and turning angle of the rat as it moved through the place field. In addition, the activity of most cells was time-locked to task-relevant approach move-ments. In the odor task, most cells fired as the rat sampled discriminative cues or when it executed specific, task-rel-

A quantitative computational theory of the operation of the CA3 system as an attractor or autoassociation network is described. Based on the proposal that CA3–CA3 autoassociative networks are important for episodic or event memory in which space is a component (place in rodents and spatial view in primates), it has been shown behaviorally that the CA3 supports spatial rapid one-trial learning and learning of arbitrary associations and pattern completion where space is a component. Consistent with the theory, single neurons in the primate CA3 respond to combinations of spatial view and object, and spatial view and reward. Furthermore, single CA3 neurons reflect the recall of a place from an object in a one-trial object-place event memory task. CA3 neurons also reflect in their firing a memory of spatial view that is retained and updated by idiothetic information to implement path integration when the spatial view is obscured. Based on the computational proposal that the dentate gyrus produces sparse representations by competitive learning and via the mossy fiber pathway forces new representations on the CA3 during learning (encoding), it has been shown behaviorally that the dentate gyrus supports spatial pattern separation during learning, and that the mossy fiber system to CA3 connections are involved in learning but not in recall. The perforant path input to CA3 is quantitatively appropriate to provide the cue for recall in CA3. The concept that the CA1 recodes information from CA3 and sets up associatively learned back-projections to neocortex to allow subsequent retrieval of information to neocortex provides a quantitative account of the large number of

cues and the associated re- behavior (Schwartz and Goldman-Rakic, 1984; Pandya sponses, perhaps providing a neural substrate for their and Barnes, 1987; Pandya and Yeterian, 1990; Barbas association. Furthermore, during learning, neural ac- and Pandya, 1991). There is some evidence that the PF tivity conveyed the direction of the animals' impending cortex is involved in conditional visuomotor learning. responses progressively earlier within each succes- Lesions of tissue around the arcuate sulcus cortex im- sive trial. The final level of activity just before the re- pair conditional visuomotor behavior, although damage sponse, however, was unaffected by learning. These included both PF area 8 and premotor area 6 (Petrides, results suggest a role for the PF cortex in learning 1982). Damage to the connections between the PF cor- arbitrary cue--response associat

... (1988), this paper compares conscious and unconscious processes across waking, nonlucid, and lucid dreams. Sleep psychology can display gross functional dissociation between perceptual and cognitive consciousness. We utilize this observation to develop models of sleep experience and dream generation. These models accommodate Hunt’s (1989) “multiplicity of dreams”, as well as the intrinsic variation of perceptual and cognitive activity during dreaming. Lucid dreams are suggested to result from the presence of a skill-based mental set, the lucid dream context, which allows voluntary interaction with the spontaneous dream process. Our view of dreaming provides an explanation of the tendency of lucid dreams to either fade or revert to nonlucid dreams. Neurobiological considerations lead us to hypothesize that, in the sleeping brain, a reversal of information flow from medial temporal lobe mnemonic structures to thalamocortical perceptual circuits imparts parameterization to dream perceptual consciousness. A consequence of our thinking is that dreaming results in a “mental recombination” of cerebral information networks, which contributes to the ability of waking consciousness to generate novel and adaptive