Abstract
Dean et al. (2007, JOV/7) found that people could trade off motor uncertainty and movement time to maximize expected gain in an economic task involving a single speeded movement. It came as a surprise that people failed to correctly allocate time between two sequential movements in a similar task (Wu et al. 2007, JOV/6). Why? Perhaps they need to learn how motor uncertainties vary with timings. Or to divide time arbitrarily. We investigated how well people could allocate time in planning a sequence of speeded movements after explicit training in varying timing.
Method: Subjects pointed sequentially to two targets on a touch screen. Successful hits resulted in monetary rewards. There were two sessions. In the practice session, subjects were trained to divide the total movement time (T = 600 ms) in specified ways (Time for the first movement, t1 = 180, 260, 340, or 420 ms). In the following test session, subjects were left to choose freely how to divide the total time. But the rewards for the two targets varied as [10 50], [10 10], or [50 10] points. Four subjects participated, three naïve.
Results: 1) In the practice session, for all subjects, the mean t1/T deviated least from the requirement in the 260/600 condition, and regressed towards 260/600 in the other timing conditions, a tendency hardly affected by practice. 2) In the test session, the three na?ve subjects used a constant time division (close to 260/600) for all three reward conditions, even if other timings they had achieved during practice led to greater expected gains. The non-na?ve subject varied timings but also failed to choose the best available ones. Our results imply that people may not be able to solve problems involving multiple motor uncertainties.