). Although this is an artificial laboratory task, it is relevant to everyday activities such as that occurring when a speaker who is delivering a talk needs to be sure to end on time. Unlike situations occurring outside the laboratory, however, arbitrary units of time were used in order to examine the effects of secondary-task demands on training and retention of a new skill, not one previously developed for producing known time intervals such as seconds and minutes. Healy et al. (2005) found circumstances under which performing an irrelevant secondary task simultaneously with time production depressed time production accuracy during training but actually enhanced time production accuracy during testing, relative to performing time production alone. Specifically, participants were trained to perform a prospective time production task, in which they were given a target number of temporal units and, after a beep, were to press the space bar when the given number of temporal units had passed. Half of the participants learned this time production task in isolation, and half learned it while simultaneously performing a difficult alphabet production secondary task, in which they were given a letter cue and were told to count backward through the alphabet by threes (e.g., m j g) until they indicated that the time interval had elapsed. Time production accuracy was much worse with the secondary task than without it during training but improved in both cases over training trials. The level of time production accuracy during subsequent testing depended on whether or not the secondary task was present during training and testing. When the secondary-task requirements were the same during training and testing, there was no loss in performance at test. When the secondary-task requirements were switched between training and testing, primary-task performance suffered at test, so that performance was no better at the start of test than at the start of training in the same condition. This was true even when training involved the difficult secondary task and testing involved no secondary task, despite the fact that the difficult task lowered primary-task performance during training. Thus, performance on time production was actually better when it was performed
Task integration in time production
University of Colorado, Boulder, ColoradoTwo experiments examined training on a prospective time production task. Participants produced intervals, expressed in fixed arbitrary units, while performing a concurrent secondary task. After a 15-min filled delay, the participants were retrained on the same tasks. These experiments tested whether the primary and secondary tasks would be integrated into a single task. In Experiment 1, the secondary-task requirements were manipulated, but the time production task was fixed. In Experiment 2, the time production task requirements were manipulated, but the secondary task was fixed. The results suggest that participants integrate primary-and secondary-task requirements.