Learning parameters for molecular memory control

Abstract

We describe a memory system which uses a molecular information processing primitive to control "reference" neurons. This system has been simulated in the context of an organism navigating in an environment. The system adapts to the environment by developing ensembles of neurons which coordinate sensory input and movement in the environment. The principal learning mechanism is based on synchronous activity of neurons. The frequency and duration of the different neural firing is critical for successful manipulation of memories developed in this system. This study suggests appropriate timing parameters for molecular processing based on the reference neuron.