dc.description.abstract | Although recent research has demonstrated training and transfer effects for task-switching, it remains to be clarified which component(s) for this important ability of executive functions was influenced by training. The current study aimed at exploring the componential process(es) that contributes to training and transfer effects in task-switching. Following the same vein, age differences were also compared.
We adopted the plus-and-minus paradigm as the training task, and implemented random switching sequence in order to study the training and transfer effects while participants did not have to actively keep track of the task to be performed. A criterion task was performed throughout pre-training, training, and post-training sessions to evaluate training effect, while a few different transfer tasks which were only performed at pre- and post-training sessions to assess near and far-transfer of training.
In Experiment 1A and 1B, participants were trained on four revised versions of univalent plus-minus switching task for ten sessions within a period of three weeks. Neither training nor transfer effect was found in both age groups. However, there was a trend of larger training effect in mixing cost for the older than the young group.
In Experiment 2A and 2B, participants were trained with the same plus-minus switching tasks for six training sessions within two weeks. The Cue-Target Interval (CTI) and task difficulty of the training task was set between the levels adopted in the near-transfer task. Yet again there was no training or transfer effect associated with age, CTI, or difficulty. We speculated that the training of univalent switching, as adopted in both Experiment 1 and 2, lacked the interference-control process in bivalent switching tasks adopted in other studies which found transfer effect.
We assessed the role of interference-control process in Experiment 3 with only young adults by using bivalent stimulus to require a higher extent of interference control for task-switching. The results showed significant training and transfer effect in the mixing cost. However, there was no far-transfer effect in the Stroop task, suggesting domain-specificity in the interference-control process.
Taken all results together, the current study showed that the training and near-transfer could be more reliably induced by training in the bivalent than in the univalent task-switching paradigm, and that the merely adding interference control may not sufficient to induce the far transfer effect in Stroop-type tasks, other working memory load may also contribute to far-transfer effects after task-switching training.
| en_US |