Time, Space, and the Hippocampus
Nestor A. Schmajuk Duke University
Hugh T. Blair Yale University
Animals build and store maps of their world in order to predict when and where different environmental events take place ( Honig, 1982). Whereas temporal maps reflect the relationships among temporal events, spatial maps reflect the relationships among spatial locations. Temporal maps are used in temporal navigation to anticipate the occurrence of environmental events. Spatial maps are used in spatial navigation to predict the location of those events. The hippocampus has been proposed to participate in either spatial ( O'Keefe & Nadel, 1978) or temporal ( Solomon, 1979) mapping.
As an alternative to exclusively spatial or temporal theories, several authors ( Moore & Stickney, 1980; Schmajuk, 1984) proposed that the hip pocampus is involved in the computation of variables common to both processes. Along that same line, Schmajuk and DiCarlo ( 1992) proposed a neural network (the S-D model) that is able to solve complex prediction problems in time and space. In the framework of the S-D model, the hippocampus modulates (a) stimulus configuration in neocortex and (b) competition in subcortical regions. Configuration refers to the combination of simple stimuli into a complex stimulus, which represents a pattern of stimuli that better predicts the future than its individual constituents. Competition refers to the selection of the stimulus best predicting the future from among different simple and complex stimuli. Schmajuk and DiCarlo ( 1992) demonstrated that the S-D model correctly simulates the performance of normal, hippocampally lesioned, and cortically lesioned