

hat do Hippocampal Neuronal Ensembles Really
Encode?
Professor Bruce McNaughton
Department of Psychology
University of Arizona
Tuscon, Arizona
 |
| Professor Seong-Gi Kim (left) leads Professor Bruce McNaughton (right) on a tour of the Center for Magnetic Resonance Research at the University of Minnesota. |
Hippocampal neurons in freely behaving rodents fire in a manner highly correlated
with the location of the animal in its environment. Given prior knowledge of
where each cell fires, the position of the animal can be reconstructed from
an ensemble of 50-100 simultaneously recorded cells with a space-time accuracy
of a few millimeters and a few hundred milliseconds. Moreover, the hippocampal
ensemble corresponding to a given location can be activated with considerable
accuracy purely on the basis of self-motion information, provided that the animal
knows its starting location (i.e., the hippocampus can be driven by a process
of path integration). Such observations provide a cornerstone of the hypothesis
that the hippocampus serves as the neural substrate of "cognitive
maps." The
map concept implies that there is metric structure in the hippocampal firing
pattern from which spatial relationships can be inferred (either by the animal
itself, or by the physiologist attempting to read the ensemble code). Two dimensional
metric structure is indeed observed under certain specific experimental conditions, but it appears to be the exception rather than the rule. Also, the collective activity patterns of hippocampal neurons in a given environment can exhibit spontaneous global rearrangements in response to internal state and behavioral context variables which do not involve changes in external landmarks. The most parsimonious interpretation of these and other observations appears to be that hippocampal activity patterns result from cooperative interactions among a large number of neurons, each of which is at best very weakly tuned to particular external features. The activity expressed by the hippocampus in a given spatial-behavioral context is probably most usefully thought of as a randomly generated tag or "index code." Spontaneous or associative reactivation of such a tag or sequence of tags may aid-via the hip- pocampo-cortical return pathways-in the coordinated retrieval of the diverse elements of a given episode from the various weakly interconnected cortical modules in which the elements are stored. Evidence for the spontaneous retrieval of hippocampal index tags and the corresponding neocortical activity patterns will be described.
|
|