Abstract
Dynamic adaptation is a key feature of brains helping to maintain the quality of their performance in the face of increasingly difficult constraints. How to achieve high-quality performance under demanding real-time conditions is an important question in the study of cognitive behaviors. Animals and humans are embedded in and constrained by their environments. Our goal is to improve the understanding of the dynamics of the interacting brain-environment system by studying human behaviors when completing constrained tasks and by modeling the observed behavior. In this article we present results of experiments with humans performing tasks on the computer under variable time and resource constraints. We compare various models of behavior generation in order to describe the observed human performance. Finally we speculate on mechanisms how chaotic neurodynamics can contribute to the generation of flexible human behaviors under constraints.