CELL ASSEMBLIES, PATTERN COMPLETION, AND THE AGING BRAIN

In the last several decades, analyses of the neurobiological sequelae of
the aging process has led to increasingly sophisticated understanding of
both the extent and the specificity of age-related changes in brain
structure and physiology. In spite of these achievements, there remains
a relatively poor understanding of which age-related biological changes
directly underlie the accompanying loss of functional capacity, and why.
During the same era, there has been a revolution in the conceptual
foundations of cognitive science. This revolution has lead to the
development of network models at various levels of abstraction which have
achieved extraordinary success in accounting for the fundamental data of
human and animal cognition, in particular, associative memory processes.
Surprisingly, there has been little attempt to apply this powerful
conceptual framework to the design of experiments that might shed light
on the manner in which neurobiological alterations contribute to
functional impairments in the aging brain. It is the intent of the
proposed research to begin to bridge the gap between these two
disciplines with experiments that exploit recent advances in
neurophysiological recording methods, permitting simultaneous recording
from large populations of neuronal elements, within a conceptual
framework provided by the theory of attractor neural networks and
autoassociative pattern completion. Specifically, we will use the method
of cross-correlation analysis to investigate the magnitude and duration
of experience-dependent changes in synaptic coupling in young and old
animals, and to test the hypotheses that synaptic loss during aging
creates conditions of instability of neuronal cell assemblies. In
addition, we will use temporary inactivation of hippocampal commissural
connections to test the theory that the age-related loss of
neurobiological resources places it near the critical point of the
graceful degradation function that characterizes the initially slow, but
subsequently catastrophic manner in which increasing loss of synaptic
connections leads to functional degradation in neural networks. Finally,
we will investigate the possibility that neural activity related to an
impending goal is maintained in the hippocampus during a delayed
conditional spatial response task, and how the quality of such goal
representations change with age.