List of Past Events
Movement planning under risk
Dr. Laurence Maloney
Tuesday, March 02, 2004, 01:00pm - 02:00pm
Psychology & Neural Science, New York University
Movement planning under risk
I'll describe a novel movement planning model based on statistical decision theory and report experiments that compare human performance with optimal performance. In these experiments subjects had to rapidly touch a target region on a computer screen. Hitting the target within a prescribed time limit gained them a monetary reward. There were also one or more penalty regions on the screen that could partially overlap the target region. Hitting these regions incurred a specified monetary penalty. In planning movements, the ideal subject had to trade off the risk of missing the reward against the risks of inadvertently hitting one or more nearby penalty regions. The risks resulted only from their own motor variability. Our subjects did alter their performance in response to changes in the penalties and their performance across experiments was not far from optimal (their winnings were typically 90% or more of what an ideal movement planner with their motor system could have won). Our results indicate that subjects effectively used an estimate of their own motor variability in planning their movements. I'll also describe an experiment in which we increased the subject's effective motor variability and tested whether s/he would correctly compensate.
Movement planning under risk is a form of decision making under risk. I'll draw parallels with the work of Kahneman & Tversky, arguing that movement planning under risk is a far more difficult problem than a simple choice among pairs of gambles and yet the human subject seems better constructed to plan movements than to make mathematically-equivalent economic choices.
Work supported by NIH EY08266, HFSP RG0109/1999-B and the DFG (Emmy-Noether-Programm)
Dr. Laurence Maloney
