Abstract: The widespread application of mobile intelligent terminals has enhanced the efficiency of individual task switching in learning and work. Studies have shown that repeated practice can effectively alleviate the impairment of judgment task switching on both reaction speed and accuracy. For example, repeated practice can reduce the response time and error rate associated with switching from Task A to Task B. Considering that both judgment tasks and memory tasks involve the engagement of executive functions during task switching processes, it remains uncertain whether the practice effect extends to memory switching tasks. Studies on retrieval practice have shown that repeated retrieval of learning content is more conducive to facilitating memory retention than repeated learning at the same time. However, whether this efficient learning method can reduce the impact of task switching on learning costs across different tasks remains to be investigated. This study aims to explore the effect of different practice methods on memory during task switching.
This study adopted a mixed experimental design of 2 (learning method: re-learning, retrieval practice) × 2 (task switching: yes, no) × 2 (test time: initial, delay), and recruited 52 college students. The results showed that the memory performance of the retrieval practice group was significantly higher than that of the re-learning group. The memory performance of the group without task switching was significantly higher than that of the mixed group with task switching. Memory performance on the initial test was significantly higher than those on the delayed test. Importantly, there was no significant difference in switch costs between the retrieval practice group and the re-learning group. However, the mixing cost (the difference between the average response time and correct rate on the repeated task between the switch group and the non-switch group) was significantly different, and the mixing cost of the relearning group was significantly higher than that of the retrieval practice group.
Based on the results of this study, we obtained three interesting findings. First, memory task switching generates learning costs. In the process of memory, task switching requires learners to spend time activating a new task set in order to complete the task successfully. That is, the activation of task switching takes up more cognitive resources and eventually produces learning costs. Secondly, compared with re-learning, retrieval practice has advantages in the learning process. The retrieval process is not only the process of recalling the content, but also the process of reconstructing the memory content, which is conducive to the improvement of the learning effect. Finally, different from the learning cost of the judgment task, this study found the phenomenon of cost reversal. Previous research on judgment tasks has found that repeated practice eliminates the mixing cost, while the switching cost still exists. However, this study finds that the mixing cost is greater than the switching cost. This discrepancy may be attributed to variations in the nature of the learning materials. Memory tasks have higher requirements for memory load, and the randomness of task presentation makes it impossible for learners to predict the form of the next task. Consequently, it becomes challenging to eliminate the impact of number recognition tasks on memory tasks.
In conclusion, this study confirms the impairment of memory caused by task switching and suggests that this may be attributed to the activation of the corresponding task set and their secondary tasks during task switching, resulting in a reduction of memory encoding resources. However, the advantages of retrieval in memory construction and search strengthen memory traces and compensate for the learning costs associated with repeated tasks in task switching.

Key words: task switching, cost, retrieval practice, memory

摘要： 移动智能终端的应用提高了个体学习过程中的任务转换频率,重复练习可缓解判断类任务转换对速度和准确性的损害,但其对记忆任务的影响尚未明确。研究采用2 （练习方式：重学,提取练习） × 2 （任务转换：是,否） × 2 （测试时间：初始,延迟）混合设计,考察不同练习方式的记忆类任务转换是否存在类似促进效应。结果发现,不同练习方式下任务转换均损害记忆准确性;初始和延迟测试（48h）中提取练习的记忆准确性均高于重学。研究表明,任务转换对记忆的损害源于任务集切换,但提取在记忆建构和搜索上的优势加强了记忆痕迹,对任务转换中的重复任务具有补偿作用。

关键词: 任务转换, 成本, 提取练习, 记忆