Healey M Karl, Kahana Michael J
Psychol Rev. 2016 Jan;123(1):23-69. doi: 10.1037/rev0000015. Epub 2015 Oct 26.
We develop a novel, computationally explicit, theory of age-related memory change within the framework of the context maintenance and retrieval (CMR2) model of memory search. We introduce a set of benchmark findings from the free recall and recognition tasks that include aspects of memory performance that show both age-related stability and decline. We test aging theories by lesioning the corresponding mechanisms in a model fit to younger adult free recall data. When effects are considered in isolation, many theories provide an adequate account, but when all effects are considered simultaneously, the existing theories fail. We develop a novel theory by fitting the full model (i.e., allowing all parameters to vary) to individual participants and comparing the distributions of parameter values for older and younger adults. This theory implicates 4 components: (a) the ability to sustain attention across an encoding episode, (b) the ability to retrieve contextual representations for use as retrieval cues, (c) the ability to monitor retrievals and reject intrusions, and (d) the level of noise in retrieval competitions. We extend CMR2 to simulate a recognition memory task using the same mechanisms the free recall model uses to reject intrusions. Without fitting any additional parameters, the 4-component theory that accounts for age differences in free recall predicts the magnitude of age differences in recognition memory accuracy. Confirming a prediction of the model, free recall intrusion rates correlate positively with recognition false alarm rates. Thus, we provide a 4-component theory of a complex pattern of age differences across 2 key laboratory tasks.
我们在记忆搜索的情境维持与检索（CMR2）模型框架内，开发了一种全新的、计算明确的与年龄相关的记忆变化理论。我们引入了一组来自自由回忆和识别任务的基准研究结果，这些结果涵盖了记忆表现的多个方面，既显示出与年龄相关的稳定性，也有衰退现象。我们通过在拟合年轻成人自由回忆数据的模型中损伤相应机制来检验衰老理论。当单独考虑各种效应时，许多理论都能给出合理的解释，但当同时考虑所有效应时，现有理论就失效了。我们通过将完整模型（即允许所有参数变化）拟合到个体参与者身上，并比较老年人和年轻人的参数值分布，开发了一种新理论。该理论涉及4个组成部分：（a）在整个编码过程中保持注意力的能力；（b）检索情境表征以用作检索线索的能力；（c）监测检索并拒绝干扰项的能力；（d）检索竞争中的噪声水平。我们扩展CMR2，使用自由回忆模型用于拒绝干扰项的相同机制来模拟识别记忆任务。在不拟合任何额外参数的情况下，解释自由回忆中年龄差异的4成分理论预测了识别记忆准确性方面年龄差异的大小。证实了该模型的一个预测，自由回忆中的干扰率与识别错误警报率呈正相关。因此，我们提供了一个关于在两项关键实验室任务中年龄差异复杂模式的4成分理论。
