摘要：Automatic acquisition of action effect associations may serve as a parsimonious account of how people acquire the basis for intentionally controlled action. However, recent research suggests that learning or the expression of action effect links might depend on whether task demands impose either a stimulus based mode of action control or an intention based action control mode. In the current study we develop a paradigm that allows the mode of action control to be varied via instructions while keeping stimuli identical. Participants were to respond to the location of a cloud of dots. Their actions were followed by predictable visual effects, either consistently congruent or incongruent with the location of the action. In Experiment 1, a displaced new cloud of random dots was presented as a spatial action effect. In Experiment 2 an arrow was presented as effect with a pointing direction congruent or incongruent to the response position. The location of the stimulus in the reference frame was easy to detect in some of the trials while the location of the cloud of dots was completely ambiguous in others. The instruction manipulation targeted the latter trials, suggesting to one group of participants to freely choose a key in a difficult trial, while asking another group to react to their spontaneous impression in the event of a difficult stimulus. In this way, we aimed at rendering actions either as stimulus driven or internally generated. By this we could investigate how effect anticipation changed with practice depending on action mode. We employed the impact of action effect compatibility on speed and choice of action as a measure for action effect anticipation. Our results suggest that action effect associations can be acquired when instructions suggest stimulus based action control or intention based action control. Instructions aiming at the mode of task processing can influence when and how action effect links influence behavior.
