Dual-Task Interference Across Practice: Reductions in Task Two Slowing

When subjects attempt to perform two speeded discrimination tasks in in rapid succession, performance on the second discrimination is often strikingly impaired. It susc previously been reported that extended practice yields only a minimal decrease in the amount of Task 2 slowing. In contrast, we found that Task 2 slowing decreases dramatically for all six subjects across 40 sessions of practice. In one experiment we demonstrate that this reduction in interference does not occur when the tasks use the same response modality. Two additional experiments investigating transfer effects demonstrate that dual-task interference remains small when a new (unpracticed) Task 2 is used, but returns to baseline when a new Task 1 is used. We conclude that the main cause of the changes that occur with extended practice is a drastic shortening of bottleneck stages in Task 1.