Impact of decision and action outcomes on subsequent decision and action behaviors

Abstract

In general, speed-accuracy tradeoff adjustments in decision-making have been studied separately from those in motor control. In the wild however, animals coordinate their decision and action, freely investing time in choosing versus moving given specific contexts. Recent behavioral studies support this view, indicating that humans can trade decision time for movement time to maximize their reward rate at the level of entire experimental sessions. Besides, it is established that choice outcomes largely impact subsequent decisions. Crucially though, whether and how a decision also influences the subsequent motor behavior, and whether and how a motor error influences the next decision is currently unknown. Here we address these questions by analyzing trial-to-trial changes of choice and motor behaviors in humans instructed to perform successive perceptual decisions expressed with reaching movement whose duration was either bounded or unconstrained in separate tasks. Results indicate that after a bad decision, subjects who were not constrained in their action duration decided more slowly and more accurately. Interestingly, they also shortened their subsequent movement duration by moving faster. Conversely, we found that movement errors not only influenced the speed and the accuracy of the following movement, but those of the decision as well. If the movement had to be slowed down, the decision that precedes that movement was accelerated, and vice versa. Together, these results indicate that from one trial to the next, humans are primarily concerned about determining a behavioral duration as a whole instead of optimizing each of the decision and action speed-accuracy trade-offs independently of each other.