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

this paper) whenever the eyes were still (to within typically 1) during the record, and calculated the firing rate together with where the monkey was looking during that record. The next record was taken immediately after the preceding one, if there was no eye movement. (The findings described in this paper were unaffected if alternatively a new record was taken only when a new eye movement was made.) The algorithm could lag its neuronal data collection a short latency later than the eye position data. (If the neuron started to respond 100 ms after the monkey moved his eyes to an effective location in space, this lag could be set to 100 ms. In practice, the lag was set for all neurons to a value of 50 ms.) From the records containing a firing rate and the place of the monkey, the head direction and the eye position, it was possible to plot diagrams and perform statistical and information theoretic analyses of the firing rate of the cell when different locations in the room were being viewed, and also in relation to eye position, place and head direction. For allocentric position, the records were binned typically into 64 bins horizontally (16 for each wall) and 16 vertically. For the information analysis, the data were further quantized into typically 16 bins, in order to provide the numbers of samples needed for the information analyses. With an average recording time for any given statistical analysis of 2.76 min, the average total number of spikes recorded from a spatial view cell was 465 spikes, leading to the average rate of a spatial view cell during these experiments of 2.8 spikes/s. (For comparison, it is of interest to note that the average firing rate of the spatial view cells described here during locomotion when the monkey is visually exploring all parts of ...

We describe here hippocampal cells that respond during whole-body motion when a monkey is moved on a remote-controlled robot-mounted platform in a cue-controlled test chamber (2 x 2 x 2 m). Some of these cells responded to linear motion, and others to axial rotation. Some of these cells responded when the same motion occurred without a view of the visual field. Such cells appeared to be driven by vestibular inputs. Other cells required a view of the visual field for their response, and these cells appeared to be driv-en by the visual motion relative to the monkey of the test chamber. Further evidence that this was the case was that some of the cells responded to rotation and linear motion of the test chamber while the monkey remained stationary. Oth-er cells responded to combinations of whole-body motion and a view of the environment.