Task-switching training was shown to improve performance not only for the trained tasks (i.e., reduced performance costs resulting from the task switches), but also for structurally similar (near transfer) or even dissimilar tasks (far transfer). However, it is still unclear whether the improvement is specific to the trained input modality or whether cognitive control occurs at an amodal processing level enabling transfer of set-shifting abilities to different input modalities. In this study, training and transfer was assessed for an auditory task-switching paradigm in which spoken words from different semantic categories were presented dichotically requiring participants to switch between two auditory categorization tasks. Cross-modal transfer of task-switching training was assessed in terms of the performance costs in a visual task-switching situation using tasks that were structurally similar to the trained tasks. The 4-day training significantly reduced the costs resulting from mixing the two auditory tasks, as compared to both an active (auditory single-task training) and a passive control group (no training). More importantly, the auditory task-switching training was also found to reduce the mixing costs for untrained visual tasks, indicating cross-modal transfer. This finding suggests that the improvement resulting from task-switching training is not specific to the trained stimulus modality, but it seems to be driven by a cognitive control mechanism operating at an amodal processing level. The training did not reveal any far-transfer effects to working memory, inhibition, or fluid intelligence, suggesting that the modality-independent enhancement of set-shifting does not generalize to other cognitive control functions.
