提取难度对困难材料提取练习效应的促进作用：来自行为和fNIRS的证据

br />The&nbsp;Retrieval Practice Effect (RPE) refers to the phenomenon whereby retrieving learned content can enhance learning and memory levels more effectively than studying repeatedly within the same amount of time. The RPE reveals the critical role of retrieval in facilitating memory retention (direct effect) and emphasizes its positive influence on metacognitive monitoring accuracy, which fosters subsequent learning and improves overall learning performance (indirect effect). Recent studies have indicated that both material difficulty and retrieval difficulty significantly influence the RPE. In this study, we investigated the impact of these factors on the direct and indirect effects of retrieval practice by manipulating the difficulty of word pairs and the level of retrieval support. Moreover,&nbsp;by incorporatingbehavioral experiments and functional near-infrared spectroscopy (fNIRS) technology, we further investigate the cognitive neuroscientific mechanisms underlying the effects of material difficulty and retrieval difficulty on the RPE.<br />Experiment 1was a behavioral study in which a 2 (word pair difficulty: easy, difficult) 2 (retrieval support: yes, no) mixed design was used to investigate the direct effects of retrieval practice on delayed testing (final test administered 2 days later). In experiment 2, an additional factor of restudy (after retrieval) was introduced and a 2 (retrieval support: yes, no) 2 (word pair difficulty: easy, difficult) 2 (restudy: yes, no) three-factor mixed design was used. This experiment&nbsp;was designed&nbsp;to further differentiated the impacts of word pair difficulty and retrieval support on both the direct and indirect effects of retrieval practice by combining behavioral measures with fNIRS technology.<br />The results showed a consistent direct effect in both experiments: the memory performance for easy word pairs was significantly higher than that for difficult word pairs, and the group without retrieval support performed significantly better on the delayed test than the group with retrieval support. The fNIRS brain imaging results&nbsp;revealed&nbsp;a significantly higher activation level in the superior temporal gyrus region (associated with retrieval difficulty) in the no-retrieval support condition than in the retrieval support condition. Moreover, learners exhibited more extensive activation in the dorsolateral prefrontal cortex (associated with cognitive load) during the retrieval of difficult word pairs than&nbsp;during the retrieval of&nbsp;easy word pairs.&nbsp;Regarding&nbsp;indirect effects, the group without retrieval support exhibited significantly superior memory performance compared with the group with retrieval support. Additionally, memory performance was significantly greater for easy word pairs than for difficult word pairs. However, a lack of retrieval support facilitated the relearning effect for difficult word pairs, leading to significant memory improvement during delayed testing.&nbsp;The fNIRS brain imaging results showed that, compared with the group with retrieval support, the group without retrieval support exhibited stronger activation of the prefrontal cortex region (associated with processing depth) during the restudy phase. Moreover, the activation of prefrontal cortex-related brain areas was reduced when easy word pairs were restudied after retrieval compared with that in initial learning phase. However, a significant enhancement in activation was observed during the restudying of difficult word pairs by the no-retrieval support group (compared with the initial learning phase).<br />The present study demonstrates that reducing retrieval support increases the difficulty of retrieval and promotes deeper processing during reencoding, which facilitates memory retention and enhances both direct and indirect effects of retrieval practice. These results provide empirical support for the retrieval effort hypothesis. Additionally, difficult materials have a weaker RPE than easy materials, but reducing retrieval support increases the retrieval difficulty and facilitating deeper processing during reencoding for difficult materials, effectively fostering the indirect effects of retrieval practice. Therefore, greater emphasis should be placed on the positive impact of retrieval-based relearning for difficult materials.&nbsp;