ADAPTIVE EXECUTIVE CONTROL - FLEXIBLE MULTIPLE-TASK PERFORMANCE WITHOUT PERVASIVE IMMUTABLE RESPONSE-SELECTION BOTTLENECKS

A new theoretical framework, the EPIC (Executive-Process/Interactive-C ontrol) architecture, provides the basis for accurate detailed computa tional models of human multiple-task performance. Contrary to the trad itional response-selection bottleneck hypothesis, EPIC's cognitive pro cessor can select responses and do other procedural operations simulta neously for multiple concurrent tasks. Using this capacity together wi th flexible executive control of peripheral perceptual-motor component s, EPIC computational models account well for various patterns of mean reaction times, systematic individual differences in multiple-task pe rformance, and influences of special training on people's task-coordin ation strategies. These diverse phenomena, and EPIC's success at model ing them, raise strong doubts about the existence of a pervasive immut able response-selection bottleneck in the human information-processing system. The present research therefore helps further characterize the nature of discrete versus continuous information processing.