Movement planning under risk differs from decision making under risk in how subjects make use of probability information

Trommershäuser, Maloney & Landy (JOSA, 2003) studied performance in tasks that were formally equivalent to decision making under risk. They found that subjects' planned movements nearly maximized expected gain, a result inconsistent with the decision making literature. Here we replicated a decision making experiment (Wu & Gonzalez, Management Science,1996) that tested whether subjects correctly use probability information in choosing between lotteries. We replicate the original experiment with the probabilities of outcomes explicitly given in the lotteries and we also replicated the experiment with each lottery translated into an equivalent motor task (“motor lottery”) where the probability of each outcome is implicit in movement uncertainty. We will describe how we measured subjects' movement uncertainty and designed an equivalent motor lottery for any given lottery. Each subject ran the implicit and explicit conditions in counterbalanced order. Task: On each trial in both conditions subjects indicated which lottery/motor lottery they preferred (2AFC). They knew that, at the end of the experiment, they would be allowed to attempt only one of their preferred explicit lotteries and one of their preferred implicit motor lotteries chosen at random and receive the outcome. Results: All subjects failed to correctly use probability information or maximize expected gain in the explicit condition, consistent with Wu & Gonzalez. Five out of eight of these subjects made choices consistent with maximizing expected gain in the implicit (motor lottery) condition. The results indicate that planning rapid movements differs qualitatively from classical decision making in how subjects make use of probability information.