Evidence for distinct steps in response preparation from a delayed response paradigm.

Abstract Task parameters still affect reaction times even when all necessary information for executing an action is presented prior to a Go signal to execute the action. Hypotheses in terms of short-term memory capacity, residual activation, and a separate motor-programming stage have been suggested to explain what can and cannot be prepared prior to a delayed Go signal. To test these hypotheses, we used a delayed response task, in which participants were to initiate a movement at onset of an imperative Go signal following the target stimulus. Across Experiments 1–3 we varied task properties including stimulus type, information uncertainty and response complexity, respectively, while controlling other factors. We also varied the time available to process the response by randomly varying the interval between onset of the target and the Go signal (i.e., the stimulus onset asynchrony, or SOA). If the preparation process is completed before initiation, the examined factor should display a strong interaction with SOA, with its effect disappearing at long SOAs. Our results showed strong, weaker, and no interaction patterns for the three factors, respectively, favoring the separate stage hypothesis, according to which response preparation is separated into steps to arrange kinematic specifications into muscle-controllable terms. Highlights • 3 experiments required a signaled response to be withheld until a go signal onset. • An effect of stimulus type on RT at short delays disappeared at long delays. • An effect of set size on RT decreased but was still significant at long delays. • An effect of response complexity on RT did not interact with delay length. • Response selection but not response initiation can be preplanned. [ABSTRACT FROM AUTHOR]