Optimizing action separation and integration through moderate practice.

Coordinating multiple actions simultaneously often results in performance costs. However, sometimes dual actions can also be easier to perform than single actions, in particular when inhibitory control is needed during single-action control. For example, an erroneous co-execution of another response or slower single- (vs. dual-) action execution may be observed. In two experiments with different extents of practice (six blocks in Experiment 1, 12 blocks in Experiment 2), we looked into two potential practice-related sources of such effects: increased action coupling with time-on-task versus initial crosstalk followed by improved separation of action representations. Participants executed single left/right saccadic eye movements, uttered single "left"/ "right" directional words, or performed both actions simultaneously based on prior cues. Results showed high saccadic co-execution error rates and vocal dual-action benefits in response times at the beginning, with a significant reduction over the course of the session. The findings support the assumption of a more pronounced separation of action representations with practice, while at the same time improvements in temporal dual-action coordination took place. The study thus highlights the optimization of action representations resulting from moderate practice.