Molecular mechanisms of associative memory and their clinical implications

Abstract

In order to study how the human brain acquires, records, and recalls the relationships that comprise the images of human memory, our laboratory initiated a research strategy more than two decades ago. The strategy began with the hypothesis that the complex patterns of human memory are constructed from numerous simple relationships that are distributed over sensory space in our experience. This hypothesis further proposed that repeatable fundamental network architectures are distributed over brain structures to create internal images of our external and internal sensory experience. Based on this hypothesis, the first element of our research strategy was to (1) identify fundamental network architectures that learn and remember simple associative relationships such as those of Pavlovian conditioned responses; (2) demonstrate that the network biophysical and biochemical mechanisms of associative learning and memory in fundamental network architectures are conserved across species as diverse as those of snails, rabbits, and other mammals; (3) demonstrate that conserved memory mechanisms are targets of pathologic involvement in a human disease characterized by memory loss such as early Alzheimer's disease; (4) and derive mathematical and logical descriptions of the functions of biological associative network architectures during learning and memory. These descriptions would then be used to design artificial neural networks that would be implemented within computer programs. Observations demonstrating the plausibility of this research strategy are presented and discussed.