Steps toward development of an integrated neurobiological model of cocaine misuse effects

We describe how a variety of neural network models might be useful in organizing the scientific data that is accumulating on the neurobehavioral effects of cocaine misuse. Implementation involving backpropagation, adaptive resonance theory, and autoassociative memory models are discussed. These techniques appear to have considerable heuristic value in terms of ordering the neurobehavioral effects of cocaine. They promise to provide a better understanding of how this drug affects such basic cognitive abilities such as executive function, attention, memory, decision making, and impulse control. Such knowledge is essential for designing treatment programs that can capitalize on the individual´s current cognitive and behavioral assets at each stage of recovery, thereby facilitating the recovery process