重复学习可以有效提高记忆成绩，但不同的重复方式对记忆的促进效果不一。其中一个经典且稳定的发现是分散学习效应：与集中的重复相比，有间隔地进行重复对记忆的提升更优。分散学习效应虽然有诸多行为和理论研究，但还存在很多争论，特别是其神经机制还并不完全清楚。近年来，研究者使用表征模式相似性分析方法发现学习阶段神经激活模式再现程度是预测重复学习效果的有效指标。这提示，记忆痕迹的再现程度可能是决定重复是否有效的关键因素。但目前为止，还没有研究考察神经激活模式机制是否能够解释分散学习效应。 本研究使用头皮脑电技术和表征相似性方法考察在人脸材料重复学习中，神经激活模式再现是否支持更优的随后记忆表现，以及不同间隔条件（集中学习/分散短间隔/分散长间隔）下神经激活模式再现程度是否存在差异。行为结果发现，分散学习下的正确再认率（分散短间隔：59.94%；分散长间隔：61.04%）优于集中学习（55.51%），重复了经典分散学习效应；脑电活动模式分析结果发现，1）右侧额区晚时间窗的神经激活模式再现程度越高，随后记忆效果越好；2）与集中学习相比，分散间隔下的神经激活模式再现程度更优。以上结果验证了神经激活模式有效再现在重复学习中作用，更重要的是，分散学习效应可能是通过提高这种再现程度来实现记忆效果的提升。此外，通过单变量分析发现，与集中学习相比，具有随后记忆效应的中央顶枕区晚时间窗的ERP在分散学习下有更大的幅度（分散短间隔：699~781ms；分散长间隔：652~750ms），这一结果为加工不足理论也提供了神经证据的支持，即分散学习下对材料的加工强度大于集中学习，因此对记忆的促进效果也更好。 综上所述，本研究通过考察神经激活模式再现在分散学习效应中的作用，验证其是有效重复学习的关键指标，同时也为分散学习效应带来了强有力的神经机制解释。除此，我们还为分散学习效应中的加工不足理论提供了新的证据支持。这些结果为进一步理解经典分散学习效应提供了更多的思路。

Repeated experiences could improve memory performance but not all repetitions are equal. Spaced learning had been consistently demonstrated to be one of effective learning ways: compared with restudying the material in immediate succession, spacing the repetitions with time would improve memory better. Though numerous theoretical accounts have been offered, the underlying neural mechanisms remain elusive. Using multi-voxel pattern analysis, researchers recently found that more precisely reinstated representations across repeated study trials were associated with better memory, suggesting that pattern reinstatements across repeated studies might be the neural basis of effective repeated learning but still needed more evidence. Examining whether better pattern reinstatement supported spaced learning effect would enhance our understanding of the neural basis of repeated learning and further our understanding of the neural mechanisms associated with spacing learning effect. Our current study applied EEG and representational similarity analysis to address those questions. Using novel faces as studying materials, we examined whether better pattern reinstatement supported better memory performance and whether there were differences between repetition lags(massed/spaced short/spaced long).Spaced learning effects were replicated in our study supported by higher recognition hit rates under spaced condition (spaced short: 59.94%; spaced long: 61.04%) than massed condition (55.51%). We found that greater spatiotemporal pattern similarity (STPS) across repetitions of the same item, which located in the right frontal electrodes occurred in the late time window after stimulus onset, predicted better memory. Importantly, compared with massed condition, that STPS was greater under the spaced condition. Those findings supported the role of pattern reinstatement in effective repeated learning and revealed that spaced learning improved memory performance by enhancing that reinstatement. In addition, we found that repeated learning under spaced condition evoked more positive-going late-time window ERP (spaced short: 699~781ms; spaced long: 652~750ms) over central parietal electrodes which also showed significant subsequent memory effect. Those results supported the deficient-processing theory. Taken together, this study found that spaced learning improved memory performance by enhancing pattern reinstatement and further supported the role of pattern reinstatement in effective learning. Besides, our results also found direct neural evidence for deficient-learning hypothesis. Those findings could advance the understanding of spacing learning effect.