Fear conditioning model predicts key temporal aspects of conditioned response production

Abstract

Five temporal aspects of the production of a conditioned response (CR) have been documented. Most of these phenomena were demonstrated using short-delay conditioning procedures, which are appropriate for eyeblink CRs. Relatively less is known about temporal aspects of CR production in longer delay conditioning procedures, which are common, for example, in studies of fear conditioning. Here, we show through computer simulations that our circuit-level model of fear conditioning predicts that these five time-domain phenomena should also be witnessed in longer delay conditioning procedures, provided that a suitable behavioral or neural probe is used. We explain how and why these phenomena emerge naturally from this general type of circuit-level model, which relies on intrinsic neuronal dynamics for time-domain computations, and we point out that the model furnishes a computational platform for Pavlov's original notion of a slowly spreading 'wave of excitation.' In accordance with Pavlov's thinking, these various temporal phenomena are captured using a contiguity-driven learning mechanism.