Stochastic optimal control and the human oculomotor system

Neuroscientists have long been interested in how e$ciently we solve probabilistic sensory problems. In order to explore analogous questions in the motor domain, we observed the eye movements of human subjects attempting to track a visual target which moved stochastically across a computer screen. The subjects' behavior was then compared to a mathematicallyderived bound on the best performance possible in such a task. The subjects were able to perform surprisingly near the optimum under the conditions examined. These results constitute an important step in determining the e$ciency of the nervous system in the context of ongoing behavior. 2001 Elsevier Science B.V. All rights reserved.