Modeling of an Operator's Performance in a Short-Term Visual Information Processing Task

A predictive model is presented that represents the human as an information processor by mathematically simulating the results of past psychological research in choice reaction time, memory storage/ retrieval, and perceptual information processing. Both discrimination and identification tasks are characterized by ideal decisions with additive Gaussian noise. In an information processing task both presented and stored information are processed using a sequential dimensional procedure to identify a presented stimulus. This technique is shown to produce a logarithmic variation in response time as a function of the number of stimuli being presented. The model's simulation of a three-dimensional visual identification task was compared with the experimental results of three subjects. The equipment used consisted of a computer-controlled discrete dot display; the dimensions used were horizontal and vertical extent and the percentage of dots present in a given area. It was found that the simulation reflected the subjects' behavior in both decision accuracy and response time for different sets of a priori probabilities.