Abstract
In motor tasks with explicit rewards and penalties, humans choose movement strategies that maximize expected gain (Trommershäuser et al., 2003, JOSA, 20, 1419). Here, two experiments were performed to explore the link between human movement planning under risk and human decision making. In both experiments, subjects were instructed to rapidly touch a target region while trying not to hit a nearby penalty region. Each target hit yielded a monetary reward; each penalty hit yielded a monetary penalty. Late responses were penalized. In the first experiment (N = 6), the time between stimulus display and the start signal for the pointing movement was varied (start signal 0 ms, 400 ms or 1000 ms after stimulus display). In the second experiment (N = 5), subjects had to rapidly point at one of two simultaneously displayed stimulus configurations, each consisting of one target and one penalty region. In some trials, the two configurations differed with respect to penalty value and spatial arrangement of the reward and penalty region. In the first experiment, subjects' movement end points and performance remained unaffected by variations in stimulus presentation time. In the second experiment, subjects' movement end points did not differ from the distribution of movement end points in trials with only a single configuration (control experiment). In trials with two configurations, four of five subjects showed a preference for the configuration with higher expected gain. Overall performance was suboptimal for three out of five (right-handed) subjects due to a higher preference for the stimulus configuration presented in the right half of the screen (81% to 72% across the overall balanced design).
Thus, human movement strategies remained stable (and optimal) across variable presentation times and for selection movements among multiple configurations. The choice among configurations with different expected gain was sub-optimal due to the constraints of the motor system.
Deutsche Forschungsgemeinschaft (Emmy-Noether-Programm, Grant TR 528/1-2)