On the mechanisms improving dual-task performance with practice

Numerous studies showed that people have difficulty performing two tasks at the same time. This difficulty is indicated by additional performance costs in dual-task situations when compared to single-task situations, i.e. dual-task costs. However, recent evidence has shown a substantial reduction of dual-task costs through practice. The finding of dual-task costs at the beginning and the reduction thereof at the end of practice indicates that this reduction must be accomplished via specific learning mechanisms. Although such mechanisms have been addressed in previous studies, the specific mechanisms contributing to practice related dual-task cost reduction remained unknown. The aim of four studies in the present work is to specify these mechanisms by applying a dual-task situation including a visual and an auditory component task (Schumacher et al., 2001). Investigations on learning mechanisms within the component tasks aimed to identify loci of stage shortening in these tasks through practice. For the first time I showed that component task processing in dual-task situations is mainly shortened through a shortening at the response selection stage. In contrast, investigations on learning mechanisms between the component tasks focussed on the acquisition of task coordination skills. Here I provided evidence that these skills are acquired during dual-task practice and there is no acquisition of these skills during single-task practice. Additionally, I demonstrated that these skills are transferable to alternative dual-task situations. There is, however, no evidence for transfer of these skills to task switching and attentional blink paradigms. In order to further specify the result of the learning mechanisms, I showed that dual-task performance in the visual but not in the auditory task is stable after practice has finished. The present findings on learning mechanisms are integrated into a model of practiced dual-task performance, the latent bottleneck model, and new assumptions in the framework of this model are discussed. und neue Annahmen im Rahmen dieses Modells diskutiert.