Citation: International Journal of Neurology, 21-22, 83-104
Abstract: A real time neural network model is described in which reinforcement helps to focus attention upon and organize learning of those environmental events and contingencies that have predicted behavioral success in the past. Computer simulations of the model reproduce properties of attentional blocking, inverted-U in learning as a function of interstimulus interval, primary and secondary excitatory and inhibitory conditioning, anticipatory condition-responses, attentional focussing by conditioned motivational feedback, and limited capacity short term memory processing. Qualitative explanations are offered of why conditioned responses extinguish when a conditioned excitor is presented alone, but do not extinguish when a conditioned inhibitor is presented alone. These explanations invoke associative learning between sensory representations and drive, or emotional, representations (in the form of conditioned reinforcer and incentive motivational learning), between sensory representations and learned expectations of future sensory events, and between sensory representations and learned motor commands. Drive representations are organized in opponent positive and negative pairs (e.g.,m fear and relief), linked together by recurrent gated dipole, or READ, circuits. Cognitive modulation of conditioning is regulated by adaptive resonance theory, or ART, circuits which control the learning and matching of expectations, and the match-contingent reset of sensory short term memory. Dendritic spines are invoked to dissociate read-in and read-out of associative learning and to thereby design a memory which does not passively decay, does not saturate, and can be actively extinguished by opponent interactions.